What’s So Special about the Developing Immune System?

Abstract

The evolution of the subdiscipline of developmental immunotoxicology (DIT) as it exists today has been shaped by significant regulatory pressures as well as key scientific advances. This review considers the role played by legislation to protect children’s health, and on the emergence of immunotoxcity and developmental immunotoxicity guidelines, as well as providing some context to the need for special attention on DIT by considering the evidence that the developing immune system may have unique susceptibilities when compared to the adult immune system. Understanding the full extent of this potential has been complicated by a paucity of data detailing the development of the immune system during critical life stages as well as by the complexities of comparisons across species. Notably, there are differences between humans and nonhuman species used in toxicity testing that include specific differences relative to the timing of the development of the immune system as well as more general anatomic differences, and these differences must be factored into the interpretation of DIT studies. Likewise, understanding how the timing of the immune development impacts on various immune parameters is critical to the design of DIT studies, parameters most extensively characterized to date in young adult animals. Other factors important to DIT, which are considered in this review, are the recognition that effects other than suppression (e.g., allergy and autoimmunity) are important; the need to improve our understanding of how to assess the potential for DIT in humans; and the role that pathology has played in DIT studies in test animals. The latter point receives special emphasis in this review because pathology evaluations have been a major component of standard nonclinical toxicology studies, and could serve an important role in studies to evaluate DIT. This possibility is very consistent with recommendations to incorporate a DIT evaluation into standard developmental and reproductive toxicology (DART) protocols. The overall objective of this review is to provide a ‘snapshot’ of the current state-of-the-science of DIT. Despite significant progress, DIT is still evolving and it is our hope that this review will advance the science.

Sohow did we get here?—Contemplating the needs and the difficulties of evaluating the developing immune system for chemical registrations and drug development, that is. What’s so special about the developing immune system? Some colleagues insisted that we use rodents and primates to assess immunotoxicity in routine repeat dose toxicity studies presently. Why wasn’t that sufficient and if it isn’t, can’t we just do the same studies on animals a little bit younger? What do you mean not all developing mammalian immune systems are the same?! From the standpoint of the developing immune system, a rat is not a dog, is not a monkey, is not a human? Well, … no. Not exactly. As with any science the more we investigate and discover, the more complexity we reveal. For toxicology, this can present numerous challenges, especially when the revelations also present real opportunities to improve human heath and, ultimately, risk assessment. And so it is with the very young subdiscipline called developmental immunotoxicology. So, just how did we get here?

In 1993, the National Research Council (NRC) published what is now a landmark report in the area of children’s heath entitled Pesticides in the Diets of Infants and Children (NRC 1993). The report addressed the question of whether existing regulatory approaches at the time for controlling pesticide residues adequately protected infants and children, because their susceptibility to toxic effects may be different than adults. The committee found both qualitative differences in toxicity (e.g., lead and altered neurological development) and quantitative differences in toxicity (due in part to differences in pharmacokinetic and pharmacodynamic processes) can occur between children and adults and that newborns may exhibit the most significant differences in sensitivity. Among several key findings and recommendations, two are of strategic importance when considering the immune system. The committee found that “… little work has been done to identify effects that develop after a long latent period or to investigate the effects of pesticide exposure on neurotoxic, immunotoxic, or endocrine responses in infants and children.” And in its recommendations regarding toxicity testing, the committee noted that particular importance should be placed on “… tests for neurotoxicity and toxicity to the developing immune and reproductive systems” as the “… extrapolation of toxicity data from adult and adolescent laboratory animals to young humans may be inaccurate.”

This seminal report was the genesis of a nearly immediate cascade of legislation and policy, and the requisite expert scientific workshops and reports discussing some of the key conclusions and recommendations made in it (reviewed in Holsapple 2002). In 1995, the U.S. Environmental Protection Agency (EPA) sponsored a workshop on endocrine disruptors, which reported, among other things, that research was needed to determine whether sensitive subpopulations (in particular the youngest population) do exist (EPA 1995). The possibility of age-related differences in susceptibility was also the topic of an International Life Sciences Institute (ILSI) Risk Sciences Institute workshop, sponsored by EPA, in 1996. Like the authors of the NRC report, the attendees of this workshop also highlighted that because so little was known about the effects of chemicals on rats 3 to 6 weeks of age, it should be considered a high priority to assess whether compounds producing immunosuppression or hypersensitivity responses in adult rodents (>6 weeks of age) produced similar effects in immature rodents. During this same time frame, the regulatory pressure to protect children’s health was increased with the U.S. Congress passage of the Food Quality Protection Act (FQPA) and the Safe Drinking Water Act (SDWA) in 1996 requiring separate risk assessments for infants and children and requiring EPA to perform studies to assess susceptibility differences, respectively. In 1997, President Clinton issued an Executive Order (no. 13045) requiring all federal agencies to make it a priority to assess differences in the susceptibility to risk in children.

EPA subsequently (1998) revised its Office of Prevention, Pesticides and Toxic Substances (OPPTS) test guideline for two-generation reproductive toxicity testing to include an assessment in F₁ and F₂ weanlings of spleen and thymus weights from one randomly selected pup per sex per litter. Though this revision was intended as a first screen to determine if additional testing was warranted, there are only limited data to support this specific approach to the identification of potential effects on the developing immune system. Also in 1999, from a workshop to identify critical windows of exposure for children’s health, it was clear that with respect to the developing immune system, the information was still scarce regarding whether chemicals impacted specific critical windows of development, but those data that were available clearly demonstrated that there are differential effects of immunotoxicants depending on the time of exposure during immunoontology (Dietert et al. 2000; Holladay and Smialowicz 2000). EPA established an expert working group in 2000 to discuss several aspects involved in the creation of a developmental immunotoxicology test guideline, including whether the state of the science supported its creation and what endpoints could be evaluated. During 2001 to 2003, workshops were held in both the United States (Holsapple 2002; Holsapple et al. 2003; Luster et al. 2003) and Europe (Richter-Reichelm et al. 2002) to further discuss the state of the science of developmental immunotoxicology (DIT). The European workshop concluded that there was sufficient evidence that differential susceptibility exists and that existing regulatory guidelines need to also consider this observation. Both the ILSI Health and Environmental Sciences Institute (HESI) workshop (sponsored by the Immunotoxicology Technical Committee with input from EPA; Holsapple 2002; Holsapple et al. 2003) and the National Institute of Environmental Health Sciences/National Institute for Occupational Safety and Health (NIEHS/NIOSH) workshop (Luster et al. 2003) considered the practical aspects of DIT testing (appropriate species/models, dosing regimen, methods and endpoints, the role of toxicokinetics, etc.), recommended a general protocol (discussed previously), and considered data gaps and future research needs. In 2002, the Food and Drug Administration (FDA) put forth guidance for industry on immunotoxicology of investigational new drugs that included a need to assess DIT of new molecular entities if immunosuppression was observed in adult animals and the patient population would include pregnant women, though no specific recommendation for how to assess DIT was made. And finally, in 2006 FDA released guidance for industry on the nonclinical safety evaluation of pediatric drug products, and this document (addressing juvenile toxicity testing guidance) recognizes that there are different endpoints in the development of the immune system, and that these responses mature at different rates. The regulatory aspects of DIT will be further considered later in this review.

Suffice it to say, albeit legislative and public pressure to protect children’s health has been significant, academic, industrial, and regulatory scientists have made a truly concerted and collaborative effort to allow good science to drive testing recommendations or requirements for developmental immunotoxicology. This is clearly because it was evident that although there were observations suggesting the developing immune system could be more susceptible to toxicants, it was just as obvious that there was much we didn’t know about the developing immune system. The remainder of this review will consider that evidence supporting the unique sensitivity of the developing immune system, the development of a testing framework including current thinking regarding functional evaluation (a cornerstone of testing in the adult animal) as well as novel information on immunopathology in the developing animal that will shape future testing protocols, and finally, the regulatory perspectives that tie science to policy.

EVIDENCE THAT THE DEVELOPING IMMUNE SYSTEM MAY BE UNIQUELY SENSITIVE TO TOXICANTS

In the context of the immune system and immune responses, initial contacts with infectious agents (i.e., important examples of environmental stressors) will lead to immunity in most cases, and consequently, encounters with the same agent later in life will lead to less severe reactions. It is known that common infectious diseases occur more often, and are usually more severe in the very young when compared to adults, and that in some cases, age-related physical or physiological differences in immune tissues or organs are responsible for the increased susceptibility to infections. In addition, increased susceptibility can be due to the relative immaturity of the immune system in the very young that prevents the host from making an adequate response to microorganisms. Neonates are particularly susceptible to infectious agents that require adult-like production of antibodies and complement to mediate phagocytosis and bacteria killing. For children who survive with primary immunodeficiency diseases (i.e., primarily those with deficiencies only in antibody production), there are increases in the frequency and severity of infections to both common and opportunistic infections. Children with less severe immunodeficiency usually have a higher incidence and severity of infections from common pathogens, such as upper respiratory infections or repeated inner ear infections, than the general age-matched population. The occurrence of asthma is also more prevalent in young children compared with adults. Although there are many intrinsic factors that influence this susceptibility, exogenous factors such as nutritional status, environment, and exposure to certain pharmaceuticals and environmental chemicals have been reported in recent years to exacerbate these immune-related effects (Akinbami and Schoendorf 2002; Smyth 2002; Timonen and Pekkanen 1997; Thurston 1997; von Ehrenstein et al. 2000; Weisglas-Kuperus et al. 2000). Because of some of these observations, one concern that has been expressed is the possibility that children may have weaker responses to some childhood vaccinations following in utero or neonatal exposure to immunosuppressive agents, and recent epidemiological evaluations of children exposed to environmental polychlorinated biphenyls (PCBs) suggests this may, in fact, occur (Weisglas-Kuperus et al. 2000, 2004; Heilmann et al. 2006).

From a toxicologic perspective, the interest in DIT has been predicated around the possibility that the immune system may exhibit greater susceptibility to chemical perturbation during ontological development that may not be detected if immune function is only evaluated in adult animals (Table 1). This greater susceptibility may be manifested as a qualitative difference, in the sense that a chemical could affect the developing immune system without affecting the adult immune system (Figure 1A ), or as a quantitative difference, in the sense that a chemical could affect the developing immune system at lower doses than the adult immune system (Figure 1B ), or as a temporal difference, in the sense that a chemical could produce a more persistent effect in younger animals than adults (Figure 1C ). Indeed, numerous examples demonstrate that the developing immune system is more sensitive than that of the adult for some immunotoxic alterations (reviewed in Dietert and Piepenbrink 2006a; Luebke et al. 2006), and the nature of the changes may be subtle in terms of routine non-functional assessment endpoints (Dietert and Piepenbrink 2006a, 2006b). This apparent sensitivity may be due in large part to novel immune maturation events. The need for rapid perinatal changes in functional balance (Th1 and Th2)—to restore the critical immune balance for protecting childhood health through the enhancement of Th1 capacity in the newborn (Holt et al. 2005; Yun and Lee 2005)—may contribute to the heightened immunotoxic sensitivity seen in early life compared to that of the adult. As a result, although effective DIT assessment should certainly draw upon the prior experience with adult-exposure immunotoxicity assessment, it is important to examine the database of known immune changes that are specific for DIT (Dietert and Piepenbrink 2006a).

There is little doubt that the potential for immunosuppression has received the most attention in regards to the developing immune system, much as it has in the adult. However, if one compares the nature of the immunotoxic changes arising from early life stage exposures to xenobiotics, it is clear that immunosuppression is only one of several health concerns (Dietert and Piepenbrink 2006b). In addition to the potential for increased susceptibility to infections (Dellaire et al. 2006) and cancer (Smith et al. 2006), additional concerns include the likelihood that some early life exposures to immunotoxicants may increase the incidence of childhood asthma and later life autoimmunity (Holladay 1999; Selgrade et al. 2006; Yeatts et al. 2006; Edwards and Cooper 2006). In fact, some developmental immunotoxicants seem capable of inducing targeted immunosuppression while at the same time elevating the risk of allergy and/or autoimmunity (Haggqvist et al. 2005).

ASSESSMENT OF DEVELOPMENTAL IMMUNOTOXICOLOGY

Immune System Development Across Species

Though it may seem that toxicological testing of the developing immune system is much like that of the adult, a deeper review of the literature reveals many more complex factors that can affect the ability (at present) to evaluate and functionally examine this dynamic system. Some of the most notable complexities are seen when comparing immune ontology across species. To begin to understand this, one needs to understand the ontology of the immune system in typical toxicologic test species.

Ontology of any mammalian organ system is a carefully orchestrated and highly regulated sequence of events. For the vertebrate immune system, these events are meticulously timed and coordinated, and begin very early in fetal life and continue through early postnatal development. The key developmental processes are the same and define the “critical windows” of development where the relative risks associated with exposure to immunotoxicants are likely to be different. Immune development across species has been extensively reviewed elsewhere (Barnett 1996; Dietert and Smialowicz 2000; Felsburg 2002; Landreth 2002; West 2002; Holsapple et al. 2003; Hendrickx et al. 2005; Dietert and Piepenbrink 2006) and is illustrated in a representative fashion in Figure 2 for humans and the three primary toxicologic test species (note: data for “rodent” are predominantly derived from the mouse and extensive immunoontologic characterization for the rat is still ongoing). Critical developmental windows for the immune system are approximated using the colored arrows in each figure.

The earliest hematopoietic progenitors are derived from uncommitted stem cells in the mesoderm within the intraembryonic splancnopleure around the heart. These cells migrate to the fetal liver and fetal spleen where they begin their differentiation into hematopoietic lineages, and then begin seeding the various tissues (monocytes) and the developing primary (thymus, bone marrow) and secondary (spleen, lymph nodes) immune organs. Lineage-restricted differentiation continues and the immune system begins to mature in its ability to respond to internal and external stimuli. During this time (until parturition), the immune system is skewed toward a Th2 bias in order to protect the pregnancy from maternal Th1-mediated rejection (Lim et al. 2000). The final stage of development is the maturation of functional responses to the adult level and the ability to generate immunologic memory.

As mentioned previously, the general processes by which the immune system develops are essentially the same across all mammals. However, as illustrated in Figure 2, the timing of these events relative to each trimester and birth are different. It is evident that the development of the immune system in the rodent is delayed relative to the human, and that unlike humans, fetal and neonatal rodents are not fully immunocompetent at birth. When considering toxicological testing results, this variance alone may result in significant differences between the response of the rodent immune system and that of the human to exposure to a developmental toxicant. Like humans, dogs and nonhuman primates have functioning immune systems at birth, but the dog is developmentally delayed in utero compared to the human, and even among primates, the monkey and human still possess developmental differences in utero.

In addition to species-related differences in developmental landmarks of the immune system, differences between the sexes as well as the placentation of the species should be considered (reviewed in Moffett and Loke 2006; Enders and Carter 2004). There is sexual dimorphism in immune response genes that begins to appear at puberty (Lamason et al. 2006). With respect to placentation, humans possess a discoid, hemochorial placenta in which maternal blood comes into direct contact with the chorion, thereby allowing materials (e.g., O2–CO2, water, inorganic ions, immunoglobulin G [IgG]) to pass between the maternal and fetal blood streams through a single vessel wall. Like humans, monkeys are hemochorial, though the attachment is bidiscoid. Although rats and mice are also considered discoid and hemochorial like humans, they are slightly different in that there are three trophoblast layers between the maternal and fetal environments. Dogs are endotheliochorial where the maternal blood is separated from the chorion by the maternal capillary epithelium. Placental differences present challenges for evaluating chemicals that are highly protein bound and may require active transport across membranes. This is evident when considering the neonatal acquisition of innate immunity (Barnett 1996). As mentioned, in humans, IgG antibodies are actively transported across the placenta into the fetus, whereas many other mammals receive antibodies only through the ingestion of colostrum. From a toxicologic perspective, some compounds may be readily transported through the placenta of one mammalian species, but not through that of the human. The hemochorial placenta is also more sensitive to conditions (or chemical exposures) that result in oxidative stress.

These differences among vertebrates do not obviate the choice of any species for DIT assessment nor do they suggest that assessment is impossible or impractical—that is likely limited only by availability of reagents. Rather, as with any aspect of toxicology, these differences need to be clearly understood and factored into the interpretation of the data obtained.

Functional Assessment of the Developing Immune System

It is necessary to emphasize at the onset that there has been no consensus regarding the best approach to assess the potential for DIT. Many investigators have approached this question by emphasizing the importance of the developmental windows, such as gestation, postnatal, weaning, and young adult (Dietert et al. 2000). More recently, investigators have relied upon the definition of developmental toxicology from the U.S. EPA (EPA 1991; Holsapple et al. 2005)—“… the study of adverse effects on the developing organism that may result from exposure prior to conception (either parent), during prenatal development, or postnatally to the time of sexual maturation”—to consider the most appropriate approach to assess DIT. Still other investigators have focused on the perinatal period—i.e., prior to and just after birth—because this window of development is known to be replete with dynamic immune changes, many of which do not occur in adults (Dietert and Piepenbrink 2006a, 2006b).

Over the last several years, a number of workshops and roundtable discussions have been organized to discuss appropriate endpoints and methods to evaluate the potential of xenobiotics to cause DIT (Holsapple 2002; Holsapple et al. 2005; Ladics et al. 2005; Luster et al. 2003). This section, in part, summarizes the main points of consensus that have emerged from these various forums. Most studies evaluating the potential developmental immunotoxicity of xenobiotics are currently based on the assessment of the immune status of adult animals following in utero exposures (Barnett 1996; Dietert et al. 2000), as the majority of assays/endpoints that have been established for measuring the effects of xenobiotics on the immune system (e.g., the antibody response to sheep red blood cells [sRBC]) have undergone extensive evaluation in adult mice (Luster et al. 1988, 1992, 1993). Validated methods for assessing the various critical windows of immune system development, as well as studies comparing prenatal, neonatal, juvenile, and adult exposures to a xenobiotic have been deemed crucial over the last 5+ years in various settings to our understanding of DIT and whether the developing offspring may be at greater risk to certain xenobiotics. In addition, some regulatory authorities (e.g., U.S. EPA) are currently considering a testing approach for DIT, whereas others have proposed triggers for testing and suggested endpoints for the evaluation (FDA 2002). Despite this attention, research regarding DIT and the appropriate methods to assess it has been limited. Direct dose comparisons for lowest observed adverse effect level (LOAEL) and no observed adverse effect level (NOAEL) involving adults versus offspring have been conducted for only a few xenobiotics (Dietert et al. 2002, 2003, 2004; reviewed in Dietert and Piepenbrink 2006; Holsapple et al. 2005; Hussain et al. 2005; Luebke et al. 2006a).

Due to the extensive evaluation and validation of several assays/endpoints that has occurred in adult rodents, various DIT forum publications have suggested that such endpoints, like the T-cell dependent antibody response (TDAR), be incorporated into a DIT protocol (Holsapple 2002; Holsapple 2005; Ladics et al. 2005; Luster et al. 2003). Either the plaque-forming cell (PFC) response (Holsapple 1995) or an enzyme-linked immunosorbent assay (ELISA) (Temple et al. 1993) on postnatal day (PND) 42 has been recommended to measure the TDAR to sRBC in a DIT protocol. Data, however, on whether assays/endpoints optimized in adult animals would be feasible or sufficiently sensitive in assessing the functionally immature immune system of neonatal or juvenile animals or whether endpoints from descriptive studies (e.g., immune organ weights or cellularity; immunopathology) may also be predictive of adverse effects are limited. Results of a study by Ladics et al. (2000) suggest that it may not be possible to measure an antibody response in preweaning rat pups due to the immature status of their immune cells, whereas data in rat weanlings (i.e., postnatal day 21 or older animals) indicate that an antibody response to sRBC of sufficient magnitude can be measured with the PFC assay. Host-resistance assays were not considered to be appropriate for a DIT screen (Holsapple 2002), because they are considered a final tier of testing and are usually only conducted when data from initial screening studies indicate immune alterations. In contrast, there was a lack of consensus on what other endpoints (e.g., phenotypic analysis of immune cells by flow cytometry, macrophage function, and natural killer cell activity) should be included in a DIT protocol, as some immune endpoints have not yet been fully validated in adults or evaluated in offspring (Luster et al. 2003). Furthermore, although data indicate that the delayed-type hypersensitivity (DTH) response can be assessed in weanling rats (Bunn et al. 2001a), data are lacking as to whether cell-mediated immune (cellular immunity) assessments in younger animals are feasible.

A question central to many discussions involves whether the inclusion of a TDAR, as well as other endpoints such as immune organ weights and immunopathology, in a DIT protocol would be adequate for evaluating DIT alterations. Although data are limited, there are some examples of developmental immunotoxicants (e.g., lead and 2,3,7,8-tetrachlorodibenzo-p-dioxin) that have been reported to affect only the cellular immunity system (e.g., the DTH response) (Miller et al. 1998; Gehrs and Smialowicz 1999; Bunn et al. 2001a, 2001b), or in the case of dexamethasone (DEX) following fetal exposure, produce a marked and persistent alteration of the DTH that was not observed in adults (Dietert et al. 2003). It is important to note that the preferential alteration of cellular immunity versus humoral immunity in the developing immune system was identified as a data gap requiring further investigation. Nevertheless, it has been recommended that a cellular immunity assay be considered in any proposed DIT protocol (Holsapple 2002; Luster et al. 2003; Holsapple et al. 2005).

For measurement of cellular immunity, a “validated” DTH or the measurement of T-cell responses to anti-CD3 antibody has been suggested (Holsapple et al. 2002; Holsapple et al. 2005; Luster et al. 2003). The DTH assay is considered by the National Toxicology Program as part of the Tier II test panel for the evaluation of cellular immunity (Luster et al. 1988). In order to avoid the use of a separate group of satellite animals when assessing both the humoral and cellular immunity responses, the DTH response in weanling and adult animals can be induced after sensitization and challenge with a T cell–dependent antigen, such as keyhole limpet hemocyanin (KLH). The DTH response is then measured as an increase in rodent footpad thickness after challenge with KLH (Bunn et al. 2001a). Of concern, however, is that a DTH assay using KLH, as put forward for DIT, may not specifically measure cellular immunity. The use of KLH may also induce antigen specific antibodies that may contribute to the measured cellular immunity response, as part of a Type III hypersensitivity response. It has therefore been recommended that the KLH DTH response be renamed to better reflect the potential contribution of antibodies to the DTH measured response (Holsapple et al. 2005). Given the latter, there appears to be a need for a validated assay that truly reflects cell-mediated immunity for assessing the potential DIT of a xenobiotic, such as the measurement of cytotoxic T-cell activity, T-cell proliferative responses to antigens (anti-CD3 antibody + interleukin [IL]-2), or a DTH assay that does not entail the use of an antibody inducing antigen (Burns-Naas et al. 2001). Interestingly, the recommendation to include assessment of cellular immunity in a comprehensive evaluation of DIT is consistent with the discussion about immune imbalance during the prenatal period (predominantly Th2) and the rapid reversal after birth, and with the observation that in several DIT studies, some examples of immunotoxicity (e.g., heavy metals, nonylphenol) were observed with a cellular immunity assay, but not the TDAR (Karrow et al. 2005; Miller et al. 1998); however, it is not consistent with the recommendation that only validated endpoints be used.

Although the majority of emphasis to date with DIT has been on the potential for xenobiotic-induced immunosuppression, exposure of the developing immune system to xenobiotics has also been postulated to contribute to an increased risk of atopy and asthma (Dietert and Piepenbrink 2006) or autoimmunity. The risk of allergy and asthma may be impacted by the ability of the xenobiotic to influence the host response to microbial stimuli both in utero and postnatally (Blumer et al. 2007). Microbial stimulation of T cells via inhalation or ingestion prenatally and early in postnatal development is believed to be a key contributor to the shift from primarily Th2 to a predominantly Th1 phenotype. Inadequate “priming” of Th1 cells may then prevent the appropriate shift and thus predisposing the child to atopy (Leonardi et al. 2007; Liu 2007).

The cytokine profile during the perinatal period tends to be skewed towards Th2 cytokines such as IL-4, IL-5, and IL-13, which are important in allergic diseases (Ota et al. 2004; Ovsyannikova et al. 2003; Protonotariou et al. 2003). In addition, the immune system of neonates and infants is deficient in producing Th1 cytokine (e.g., interferon gamma [IFNγ]; IL-12) responses (Holt et al. 2003; Itazawa et al. 2003; Rowe et al. 2001). As the immune system matures, however, there is a move towards increased Th1 immunity due in part to increased exposure to infectious agents and a more balanced Th1/Th2 phenotype (Krampera et al. 1999). It has been hypothesized that xenobiotics that interfere with the development of a Th1/Th2 balance may result in an increase in certain allergies (e.g., childhood asthma) or an increased susceptibility to infections (Fahy et al. 2002; Holt et al. 2003). Importantly, these cytokine responses should not be considered all or none, but rather degrees of Th1/Th2 balance. Additional studies, however, are needed to confirm these hypotheses.

The measurement of specific cytokines or the increase or decrease in the expression of certain genes during the various ‘windows’ of development of the immune system following xenobiotic exposure may prove useful for discriminating immune status or determining disease susceptibility (Karpuzoglu-Sahin et al. 2001). However, there are a number of issues which currently limit the use of cytokines as DIT biomarkers: (1) a lack of baseline data on normal levels of cytokines occurring in immune tissues and organs during the various windows of immune system development; (2) lack of standardization of both in vivo and in vitro methods for evaluating cytokine levels; and (3) lack of understanding of the time- and dose-dependent kinetics of cytokine levels to known DIT xenobiotics in various immune tissues and organs. Likewise, there are similar limitations associated with using genomics to investigate the potential immunotoxicity of xenobiotics (Burns-Naas et al. 2006; Luebke et al. 2006). A recent workshop on genomics and immunotoxicology concluded that in addition to standardizing the reporting structure of -omics data, it was important to establish public databases containing baseline genomics data for immune tissues/organs of rodents. Such a database(s) is important in order to observe significant alterations of the transcriptomes of xenobiotic exposed neonatal and adult animals (Luebke et al. 2006).

Recent workshops and forums have also suggested that when possible, DIT assays/endpoints be added onto existing developmental and reproductive toxicology protocols (Holsapple et al. 2005; Ladics et al. 2005). Such combined–endpoint studies were encouraged for preliminary screening of new (i.e., not previously tested) xenobiotics, rather than for substances for which extensive toxicological data have already been generated. It was noted that the interpretation of such stand-alone studies could be improved if all potential xenobiotic-mediated effects could be integrated into a single study, instead of interpreting data from several studies where exposures or observed effects may have been different.

Clearly, one area where data are lacking is in regard to developmental immunotoxicity in humans. Luster et al. (2005) described a number of endpoints or biomarkers that may be of use for epidemiological studies involving DIT. The endpoints discussed by the authors included T-cell receptor rearrangement excision circles (TRECs); cytokine measurements; immunophenotyping, serum immunoglobulin measurements; and the quantification of the immune response to childhood vaccines. Although considered biologically relevant, TRECs and cytokine measurements were deemed to need further evaluation and validation prior to their use, whereas immunophenotyping and measurement of serum immunoglobulin levels were not considered to be very sensitive. In contrast, the authors suggest that measurement of the immune response to vaccines (van Loveren et al. 2001), such as hepatitis B (Kiecolt-Glaser et al. 2002) or diphtheria-tetanus-pertussis (DTP; Swartz et al. 2003), may provide a good indicator for DIT. Additional studies, however, are needed to evaluate the utility and sensitivity of such vaccines for assessing DIT.

And finally, immunopathology is an endpoint identified as being easily incorporated into existing developmental and reproductive toxicology protocols (Holsapple et al. 2005; Ladics et al. 2005). If immunotoxic effects were observed in adult subchronic studies, it has been suggested that histopathology of immune organs be included in the protocol of a reproductive toxicology study. Presently, neither EPA nor FDA require histopathology of immune organs in guideline-driven reproductive studies. There has been, however, disagreement regarding what role pathology should play in assessing DIT, and this will be considered in more detail in the following section. Routine histopathology as a stand-alone endpoint was deemed by some not to be sensitive enough to independently and consistently characterize immunotoxic effects, particularly in the developing immune system, compared to those endpoints that assess immune function (Germolec et al. 2004). In contrast, others indicated that immunopathology was an appropriate screen for immunotoxic effects (Haley et al. 2005). In a number of recent DIT investigations, histopathological examination was found to frequently explain observed functional alterations; however, there are examples where morphometric histopathologic findings do not predict DIT functional impairment (Hussain et al. 2005). It can be argued that histopathology assessment can be helpful in providing a more comprehensive picture of DIT alterations, but that histopathology alone would not substitute for functional immune tests in a primary DIT screen. This assessment of the potential value of histopathology should be made in the context of recognizing that most of the studies to date have not utilized the full spectrum of available histopathological techniques and that there is a critical need to standardize age and procedures for collection, fixation, and staining tissue samples in order to maximize the collection of valuable data from routine toxicology studies (Haley 2003; Holsapple et al. 2003). Moreover, as in the case of validation of immune assays for DIT, one concern to be addressed is that historically, pathologists have far less experience with fetal and neonatal rat immune histology than with that of the adult rat.

In addition to a need for standardization of procedures for immunopathology, standardization of age is also important. In the situations where one might consider assessment of immunopathology in the context of routine developmental and reproductive toxicology (DART) studies, no specific recommendations have been made regarding the age of the animals for immunopathology of the developing immune system, though animals aged PNDs 4, 22, and 42 have been suggested. The selection of age is important, however, because implantation of the concepti in rodents (the most frequently used test species) can vary by up to 2 days. This suggests that there may be more histologic variability in early aged pups (e.g., PND 4) compared with weanlings (i.e., PND 22) and therefore true chemically mediated effects may be more difficult to distinguish among groups.

To See, or Not to See: The Role of Pathology

Pathology evaluations have been a major endpoint of standard nonclinical toxicology tests, and could serve as a starting point for evaluation of DIT studies. The pathology evaluations of typical nonclinical toxicology protocols are designed to provide a broad-based mechanism for detection of test article–related alterations in structure or function of all major organs and tissues, with no particular emphasis on individual organs or systems. Standard protocols may be modified based on route of administration of test articles, specific questions that have arisen in previous studies of the same or similar test articles, therapeutic applications of test articles, or other factors. Standard pathology evaluations should be viewed as screening tools rather than definitive investigations.

The type of pathology evaluations used in standard nonclinical toxicology studies may be applied in developmental immunotoxicology studies, but some modifications will probably be required due to the size and age of the animals, as well as the special features of the immune system. Standard pathology evaluations are considered adequate for detection of test article–related alterations in many organ systems, but have been found inadequate for detection of chemically mediated changes in other organ systems. Additional testing procedures have been implemented when standard pathology evaluations are found to be inadequate. Examples of additional testing procedures that have been implemented when standard pathology evaluations are found to be inadequate would include QT interval prolongation studies for detection of cardiac conduction abnormalities and morphometric analysis of brains in developmental neurotoxicity studies. Additional testing procedures employed in DIT studies may involve pathology evaluations such as immunohistochemistry (Ward et al. 2007), flow cytometry, or enhanced immunohistopathology, or may require input from entirely different disciplines that are not within the scope of pathology evaluations as they are currently defined.

Although it is possible that standard pathology evaluations may detect the structural alterations associated with immunosuppression, those evaluations are unlikely to detect misdirected or enhanced immune responses. Testing methods in addition to standard pathology evaluations will almost certainly be required for detection of these latter types of immunotoxicants (Dietert and Holsapple 2006).

Pathology evaluations for purposes of this discussion are subdivided into (1) necropsy examination, (2) organ weight evaluation, (3) clinical pathology, (4) histopathology, and (5) special procedures. Following is a brief discussion of these individual evaluations as they relate to the detection of immunotoxicants in adult animals or animals from DIT studies.

Necropsy Examination

Animals from nonclinical toxicology studies are sacrificed near the termination of dosing and subjected to a thorough gross necropsy examination. Subsets of animals may be sacrificed at one or more interim time points prior to the cessation of dosing in order to determine the progression of test article–related effects, and at one or more postdosing (recovery) time points to determine reversibility of test article–related effects.

The majority of pathologic alterations are detected by histopathologic examination and organ weight analysis, thus those facets of the pathologic examination receive the greatest attention. However, some alterations are most amenable to detection by gross inspection. Among these are discolorations that may not persist through histologic processing, distention of organs that may revert to normal upon fixation, and accumulation of fluids or gases that may be removed during histologic processing. Careful examination during necropsy, meticulous recording of observations, group-related tabulation of observations, and methodical tracking of gross lesions through histologic processing and histopathologic examination are required before macroscopic alterations can serve as a meaningful basis for interpretation. Gross necropsy observations would be expected to serve a similar role in DIT or adult animal studies. Necropsy examination of the small specimens from DIT studies may be facilitated by the use of magnifying loupes and dissection microscopes, among other things.

Table 2 presents immunologically important organs that are collected during a standard non-clinical toxicology study. Preservation of a similar list of tissues would be necessary for thorough pathology evaluations in DIT studies.

Organ Weight Evaluation

Alteration (reduction) in the weight of lymphoid organs is commonly seen following administration of immunosuppressants, but there is no corresponding alteration in lymphoid organ weight that is consistently associated with enhanced or misdirected immune responsiveness. Interpretation of reduced immune system organ weights is complicated by the possibility of glucocorticoid release associated with stress, which can result in a decrease in the weight of immune system organs in some species. Organ weight interpretation is further complicated by treatment-related alterations in body weight, which may or may not be due to changes in feed consumption. Some organ weights decline synchronously with body weight, other organ weights are conserved despite alterations in body weight, and the weight of other organs is influenced to an intermediate degree by alterations in body weight. Many nonclinical toxicology studies are conducted on young, actively growing animals. Changes in the weight of individual organs may present a different pattern in animals that actually lose weight, as opposed to animals that fail to gain weight during the course of the study. Brain weight tends to be conserved despite a decline in body weight or failure to gain body weight, thus brain weight is commonly used as a reference for comparison of other organ weights. If the true significance of organ weight alterations is to be accurately determined, organ weight changes must be evaluated as absolute organ weights, relative to body weight and relative to brain weight (Schwartz et al. 1973; Scharer 1977). Careful attention should be given to feed consumption or nursing data as well as any macroscopic, clinical laboratory, or histopathologic evidence of spontaneous disease or other factors that could indicate concurrent disease or a stress response in individual animals. Organ weight analysis in DIT studies is even more difficult due to the small size of the organs and the relative variation that is introduced by technical factors such as presence of extraneous tissues or fluids on the specimens as they are weighed.

Table 3 presents immunologically important organ weights that are collected during a standard nonclinical toxicology study. A similar list of organ weights would be necessary for thorough evaluation of the immune system in DIT studies.

Clinical Pathology

The clinical pathology testing performed in a typical non-clinical toxicology study includes total white cell count and differential leukocyte count as well as determination of serum albumin, globulin, total protein, and albumin:globulin (A:G) ratio. Substantial reductions in white cell counts, particularly lymphocytes, are commonly seen following administration of immunosuppressant chemicals. Concurrent reduction in circulating granulocyte populations may indicate a pancytopenic effect on bone marrow. Increased white cell counts, commonly including a marked increase in granulocytes, is a common manifestation of inflammatory disease processes. A specific increase in eosinophil populations may indicate a hypersensitivity reaction or infestation by metazoan parasites. Altered A:G ratio may indicate an alteration in immunoglobulin levels, but albumin levels should be carefully reviewed to insure that the altered A:G ratio is indeed due to an alteration in globulin levels, as opposed to an artifact resulting from altered albumin levels. Albumin is a ‘negative acute phase protein,’ i.e., albumin production by the liver is decreased in response to cytokines released during a systemic inflammatory response (Matic et al. 2004; Sevaljevic et al. 1989).

Sample size limitations in DIT studies may hinder a complete clinical pathology evaluation of the scope that is performed in standard non-clinical toxicology studies. In this context, evaluation of hematology parameters would be more useful than evaluation of clinical chemistry parameters in the detection of immunotoxic alterations, thus hematology evaluations should be given a priority when sample size is limited. Modern instrumentation allows complete hematology evaluations of samples as small as 200 to 300 μl, which are available from neonatal rats.

Histopathology

Histologic changes associated with altered immune responsiveness may consist of (a) structural alterations in the primary or secondary organs of the immune system, and/or (b) secondary evidence of altered immune system function with resultant changes in a variety of organs. Normal and abnormal histologic features of the major organs of the immune system are reviewed in a recent monograph (Cesta 2007a, 2007b; Elmore 2007; Kuper 2007; Maronpot 2007; Pearse 2007a, 2007b; Suttie 2007; Travlos 2007a, 2007b; Willard-Mack 2007).

Structural Alterations in Organs of the Immune System. Histologic examination of specimens from nonclinical toxicology studies may reveal reductions in cell populations associated with reduced immune responsiveness but would not reveal increased or misdirected immune responsiveness unless the immune system alteration results in structural damage to another organ or tissue. Reduction in the population of mature lymphocytes in the thymic cortex is a common manifestation of immunosuppression. The histologic presentation may consist of a reduction in the ratio of cortical to medullary areas, commonly described as lymphoid atrophy or lymphoid depletion. Lymphoid atrophy/depletion may result from death of cortical lymphocytes, or a reduction in the rate of proliferation of cortical lymphocytes. Given the rapid rate of cell proliferation in the thymus, any interference with cortical lymphocyte proliferation quickly results in a pronounced reduction in the overall cellularity of the thymic cortex. Reduction in thymic cortical cellularity related to test article administration must be distinguished from normal involution of the thymus, which is histologically similar. Review of the age of individual animals may be helpful in making the latter distinction, though experience suggests there is great variation in the age of onset of senescent changes in the thymus of the common laboratory animals.

Depending on the time elapsed between thymic insult and tissue collection, one may encounter direct histologic evidence of active necrosis or apoptosis of thymic cortical lymphocytes (Besteman et al. 2005). Although these changes probably progress to a stage identifiable as lymphoid atrophy/depletion, thus representing a different stage in the same pathologic process, the active necrosis/apoptosis lesions should be identified by some distinctive term that allows them to be separated from the inactive terminal lesions.

Some degree of thymic lymphocyte apoptosis is expected due to negative and positive selection processes. Distinguishing this ‘background’ level of cell death from test article–related changes can be challenging when the test article effects are subtle. Application of rigorous diagnostic criteria, including severity scoring criteria, is critically important in making these distinctions. As discussed below, digital image analysis performed on sections subjected to immunohistochemical staining for apoptosis markers may provide a more definitive indication of the level of thymic apoptosis.

Histologic sections from non-clinical toxicology studies are typically 5 to 6 μm in thickness, and are stained with hematoxylin and eosin (H&E). Such sections are adequate for their intended purpose, but often prove to be too thick for critical histologic evaluation of lymphoid tissues. For example, histologic sections of standard thickness may not reveal sufficient detail in a densely cellular tissue such as the thymic cortex. When possible, histologic sections of lymphoid tissues should be prepared at 2 to 3 μm thickness.

Immunotoxicants may result in subtle structural alterations in the immune systems that are not readily classified by standard histopathologic evaluation, which is designed to detect and assign subjective severity scores to ‘lesions’ (Kuper et al. 1995, 2000; Hinton 2000). Immunotoxicant-associated structural alterations may fall within the broad range of ‘histologically normal,’ thus would not be detected by standard histopathology procedures. Guidelines for enhanced histopathologic evaluations of the immune system organs have been proposed in an attempt to address the deficiencies of routine histopathology (Haley et al. 2005). The enhanced evaluation consists of subdividing organs of the immune system into compartments and applying numerical scores to various structural attributes of the compartments, regardless of whether those attributes are judged to be ‘normal’ or ‘abnormal’ (Elmore 2007aElmore 2000b, 2000c, 2000d, 2000e, 2000f; Maronpot 2007). An example of such an enhanced histopathologic evaluation is depicted in Table 4.

Rats have historically been the primary rodent species used in nonclinical toxicology studies, whereas mice have traditionally been the model of choice for immunology studies. As a result, more information is available regarding the immune system of mice than of rats. However, due to the widespread use and extensive background toxicology data that are available on rats, it has been suggested that rats should be used for DIT studies (Holsapple et al. 2005). It has also been proposed that a DIT component could be incorporated into currently accepted DART protocols (Ladics et al. 2005).

There are questions as to whether either the standard or enhanced pathology evaluations that are applied to adult animals will produce reliable information when applied to juvenile animals. The histologic features of immune system organs in adult laboratory animals are well known (Dunn 1954; Haley 2003), but those features are not as well defined in neonatal animals. If histopathology is used as a major component of the screening process in DIT studies, it is imperative that we have in-depth information regarding the histologic features of the developing immune system organs in late gestational and neonatal laboratory rats.

Preliminary studies suggest there is a remarkable progression in the histologic appearance of fetal organs, including organs of the immune system, that takes place between gestation days (GDs) 15 and 20. At GD 15, the only visceral organ with histologic similarity to the mature counterpart is the liver, which consists of approximately equal hepatocellular and hematopoietic elements (Figure 3A ). However, by GD 20, the rat fetus has easily recognized internal organs, including thymus, spleen, and mesenteric lymph nodes (Figure 3B ).

At GD 20, the fetal thymus has distinct corticomedullary relationships (Figure 4A), though the organ lacks the dense cellularity seen at PND 22 (Figure 4B ). At PND 22, the thymic cortex has such intense cellular proliferation that virtually all cortical cells are stained by proliferation markers such as Ki-67 (Figure 4C ). The dynamic nature of the thymus, which persists into early adulthood, is the basis for the rapid degenerative and regenerative changes seen in response to some immunotoxicants.

At GD 20, the spleen is easily recognized but lacks adult-type internal structures such as periartiolar lymphoid sheaths (PALs), lymphoid follicles, or germinal centers (Figure 5A ). By PND 22, the spleen has the typical complement of PALs, with indistinct lymphoid follicles and few germinal centers (Figure 5B ).

Mesenteric lymph nodes are discernible at GD 20, but lack lymphoid follicles or other evidence of immune responsiveness (Figure 6A ). By PND 22, the mesenteric lymph nodes have distinct primary follicles (Figure 6B ), but secondary follicles and germinal centers are not apparent.

Development of secondary lymphoid elements such as bronchus- and gut-associated lymphoid tissue (BALT and GALT, respectively) seems to follow exposure of the newborn animal to the external environment. Sections of lung and intestine from fetal rat at GD 20 had no discernible BALT (Figure 7A ) or GALT (Figure 8A ), respectively, whereas those from a rat at PND 22 had small but distinct aggregations of BALT (Figure 7B) and GALT (Figure 8B ). Though present, GALT and BALT in early neonatal rats appeared to be relatively inactive, as opposed to the overt proliferative activity that is seen in these secondary lymphoid structures in young adult rats.

The bone marrow of fetal rats at GD 20 has a population of hematopoietic cells (Figure 9A ), but lacks the intense cellularity seen in the bone marrow by PND 22 (Figure 9B ). Histologic examination of the bone marrow should be possible by PND 22, though it is generally acknowledged that histologic examination of bone marrow can provide only limited information. Examination of properly stained cytologic smears provides more detailed information regarding bone marrow cell populations.

The liver is not considered to be a primary or secondary organ of the immune system, but is known to participate in immune responses (Parker and Picut 2005). The liver is recognizable as a distinct organ at GD 15, at which time it has an abundance of hematopoietic activity (Figure 10A ). By GD 20, the hematopoietic activity in the liver has subsided to some extent (Figure 10B ), presumably due to the migration of hematopoietic activity to the bone marrow. By PND 22, the liver has only traces of hematopoietic activity (Figure 10C ).

In summary, early studies suggest there is little potential for histopathologic detection of chemically mediated alterations in the developing immune system of laboratory rats at or prior to GD 15. Major perturbations (e.g., failure in the formation of immune system organs) could be detected by standard histopathologic examination by GD 20. The immune system of the untreated, developing rat is anatomically intact by PND 22, perhaps earlier, but histologic features at PND 22 suggest relative inactivity (lack of stimulation by exogenous antigens). Further studies are underway to determine a more precise time sequence in the development of immune system organs in rats.

Some attention should be given to the placenta as a possible site of chemically mediated immunomodulation. Placentas are classified by shape (diffuse, discoid, zonary, or cotelydonary) or degree of separation between maternal and fetal blood (epitheliochorial, syndesmochorial, endotheliochorial, or hemochorial) (Bjorkman and Dantzer 1987). Close approximation of maternal and fetal blood allows placental transfer of maternal antibodies to the fetus. Species with a greater separation between maternal and fetal blood rely on colostral transfer of maternal antibodies. The placenta of humans and rodents is hemochorial, which results in the closest possible approximation of fetal and maternal blood (Figure 11) and resultant placental transfer of maternal antibodies to the fetus. Chemically induced alterations in the placenta could influence this antibody transfer. It should be noted that presence of the hemochorial type of placentation does not completely eliminate the role of colostral antibody transfer.

Secondary Evidence of Immune System Dysfunction. Although most modern, accredited laboratory animal facilities are maintained to very strict standards of hygiene and animal husbandry, nonclinical toxicology facilities typically are not germ-free. Chemically mediated immunosuppression may result in increased susceptibility to pathogens, resulting in outbreaks of specific diseases or a generalized increase in incidence of various infectious diseases. Outbreaks of specific diseases can present as a myriad of viral, bacterial, protozoan, mycotic, or parasitic diseases, some of which are transmitted between laboratory animals and animal handlers. Outbreaks of specific diseases are easily recognized, but recognition of a generalized increase in incidental infectious diseases or inflammatory lesions can be more problematic. In the latter situations, histologic alterations may involve a number of different organs, with no single organ being involved to a degree that results in an unequivocal treatment-related incidence pattern.

Histologic lesions associated with reduced immune system function are commonly located in organ systems that have direct contact with the external environment, principally the integumentary, respiratory, gastrointestinal, and urinary systems. Immunosuppressant-related lesions of the respiratory and gastrointestinal system can be histologically spectacular and clinically catastrophic. However, subtle increases in the incidence and/or severity of ‘background’ lesions such as nonspecific prostatitis in rats or exacerbations of Demodex canis infection in dogs may be equally important as indicators of immune system modulation. These effects may involve changes in severity rather than, or in addition to, incidence, thus it is critical that all observations should be graded with regard to severity or quantified in some appropriate manner.

Detection of a chemically mediated increase in the incidence and/or severity of spontaneous disease processes must involve clinical observations and clinical pathology tests as well as histopathology, and should involve a ‘whole animal’ approach. Simply tabulating histopathologic lesions or clinical signs, or performing statistical analysis on clinical laboratory data, in the absence of this ‘whole animal’ approach may result in failure to detect, or misidentification of, many chemically mediated immune system alterations.

Regulatory Perspectives

Environmental Chemicals

For environmental chemicals (including pesticides), functional assessment of the immune system is ascertained only in young adulthood in specific immunotoxicity studies. As described in Kimmel et al. (2005), currently there are only two testing guidelines that evaluate the developing immune system: prenatal developmental toxicity (OPPTS 870.3700) and the two-generation reproduction study (OPPTS 870.3800). EPA had revised its OPPTS test guideline for two-generation reproductive toxicity testing in 1998 to include an assessment in F₁ and F₂ weanlings (~PND 21) of spleen and thymus weights from one randomly selected pup per sex per litter, a revision intended as a first screen to determine if additional testing was warranted. Structure (macroscopic pathology) of the spleen and thymus is examined at the end of gestation in the prenatal developmental toxicity study and in early adulthood in the reproductive toxicity study, following developmental exposures to the chemical. Though these endpoints could potentially provide triggers for additional DIT testing, it is clear that these protocols may miss adverse events occurring during the juvenile and early adolescent stages by not specifically evaluating immunopathology and functionality (EPA 2002a). Additionally, latent effects could also be missed. Thus, EPA established an internal expert working group in 2000 to discuss several aspects involved in the creation of a developmental immunotoxicology test guideline, including whether the state of the science supported its creation and endpoints that could be evaluated. More recently, in light of the recommendations from the ILSI HESI Agricultural Chemical Safety Assessment (ACSA) Technical Committee (Cooper et al. 2006) and the encouragement by the EPA to combine endpoints into existing protocols where feasible, an extended one generation DART protocol is under discussion at the EPA as a possible protocol that would evaluate multiple developmental endpoints and could, depending on outcome, eliminate the need for specialized testing (e.g., DIT, developmental neurotoxicity [DNT], etc.) later in development. This protocol is discussed further in the next section of this review. Meanwhile, EPA has approached DIT testing on a chemical-specific basis, reliant upon a weight-of-evidence review of the existing data base, which may include an assessment of immune function in young adult rats. DIT testing in support of chemical regulation has been recommended by the Agency in only a handful of cases to date. Because no specific DIT guideline has been developed, EPA relies upon recommendations from the published literature for guidance on protocol specifics.

Pharmaceuticals

Consistent with the approach taken for assessment of immunotoxicity in adult animals, which provides for a case-by-case determination based on consideration of a weight of evidence review of available data and key potential triggers (ICH 2005), global regulatory agencies have not routinely required assessment of DIT during the development of therapeutic agents. Rather, these assessments are driven by “triggers,” or a “cause for concern,” which is derived from a weight of evidence evaluation of data obtained in first-in-human (FIH)-enabling repeat-dose toxicity studies in adult animals. Studies in adult animals that show the immune system as a potential target organ may suggest the need for an assessment of DIT, as the data to date show that all chemicals that are immunotoxic in adults to be immunotoxicants in the developing immune system when evaluated, often with differential sensitivity between the two age groups. Potential signs of immunotoxicity could be structural or functional. Other possible triggers are the intended patient population and the potential for neonatal exposure. For example, drugs used to treat immunocompromised individuals, including children, may warrant consideration for DIT testing; and if pregnant or lactating women are to be treated, and the drug may cross either the placenta or into breast milk (or when this information is unknown), one may wish to consider the need for DIT testing. Though this seems to leave the door open for quite a bit of interpretation, what is clear is that decisions to conduct DIT studies should be based on sound science and should aid in the overall risk evaluation of the proposed new drug. For example, if it is known that in utero exposure is very low or nonexistent, then performing a DIT assessment may not provide information that would significantly aid in assessing the overall risk of the drug. Additionally, DIT studies should not be routinely performed, particularly in nonrodents, simply to minimize regulatory risk for the registration.

Global regulatory authorities generally concur that evaluation of potential adverse effects of human pharmaceuticals on the immune system should be incorporated into the standard drug development process (ICH 2005), though this specifically applies to the adult animal. In the United States, the FDA first addressed the potential need for DIT testing in their guidance for industry on the nonclinical evaluation of immunotoxicity for new chemicals (FDA 2002). In it the Agency notes that a DIT assessment should be considered if the drug is “expected to be used in pregnant women and has been shown to induce immunosuppression in adults. …” The European Committee for Human Medicinal Products (CHMP) has taken a similar approach in its most recent draft guidance on nonclinical testing to support pediatric drug development (EMEA 2005). CHMP recognizes that major developmental differences exist between the immune systems of human neonates/infants and adults, and considers that these developmental differences are generally apparent until age 12. CHMP also indicates that pre- and postnatal exposure can potentially result in all types of immunotoxicity in the offspring, including immunosuppression, hypersensitivity, allergy, and autoimmune disease. Juvenile studies have been suggested to be performed “on a case-by-case basis and only after a careful consideration of the available data and the age and duration of treatment of the intended pediatric population,” but that “if any of the major functional systems are shown to be potential targets, either from human or from nonclinical studies, studies in juvenile animals should be considered.” Further, “immunotoxicity studies are only required if the chemical/pharmacological class of compound or previous studies in humans or animals gives cause for concern for the developing immune system.”

FDA’s juvenile testing guidance (FDA 2006) also recognizes the immune system as one that continues to mature after birth (until approximately 5 to 12 years of age) and may well be a novel target for drug-induced toxicity. Like its 2002 guidance on immunotoxicity testing, and the CHMP draft guidance on nonclinical juvenile testing, FDA considers that DIT evaluations may be warranted on a case-by-case basis with consideration given to “(1) the intended or likely use of the drug in children; (2) the timing of dosing in relation to phases of growth and development in pediatric populations and juvenile animals; (3) the potential differences in pharmacological and toxicological profiles between mature and immature systems; and (4) any established temporal developmental differences in animals relative to pediatric populations.” FDA also considers that “the greatest concern is with chronic, long-term therapy,” and thus “the duration of anticipated treatment of the pediatric population should be considered in relation to the duration of developmentally sensitive phases.”

Development of a Testing Framework for DIT

From a regulatory perspective, the interest in DIT has been predicated around the possibility that the developing immune system may exhibit greater susceptibility to chemical perturbation than young adults, and that guideline immunotoxicology studies conducted exclusively in young adult animals would not detect this greater susceptibility. Although immunotoxicology has evolved to the point where testing guidelines (young adult animals) exist within many regulatory frameworks, the discussion to incorporate appropriate experimental approaches and assays available to assess DIT is clearly still quite limited, in spite of an impressive level of activity in this area over the last few years. Regulatory concerns about potential risks to the developing immune system from exposures to pharmaceutical or environmental chemicals could lead to the decision to require a DIT study for a specific test substance. The discussion over the need for a special (dedicated) DIT protocol must consider the premise that a well-designed and well-executed guideline immunotoxicity study in young adult animals would miss the potential for unique susceptibility in the young. So what would a ‘catch-all’ protocol look like?

Up until recently most studies conducted to evaluate the DIT effects of drugs or chemicals have utilized an exposure regimen where dams are exposed to the chemical during gestation and/or lactation, and the immune systems of the offspring were evaluated as young adults. Importantly, the specific conditions for exposure vary from study to study and have involved mostly acute, high-dose exposure at one or more critical developmental windows. This is an important consideration when interpreting the data for human risk assessment or generalizing DIT. Notably, extensive characterization of effects across gestation or between gestation and lactation has not generally been performed. The period of nonexposure prior to immune evaluation may make this protocol more applicable for the evaluation of persistent, and perhaps permanent, immune effects, but leaves open the possibility that biologically significant effects may occur at earlier times and could be missed due to recovery during the period of nonexposure. In independent workshops, immunotoxicologists and developmental scientists agreed that the best approach to DIT testing may be to address all critical windows at once, evaluating immune function at PNDs 42 to 56, and then perform assessments of specific windows if warranted from the results (ITC and Sandler 2002; Holsapple 2002; Luster et al. 2003; Holsapple et al. 2005). This approach allows no opportunity for recovery and is illustrated in Figure 12A . Initiation of direct dosing to offspring would be determined by whether it was known if the chemical was excreted into breast milk. An alternative DIT protocol is presented in Figure 12B whereby the study is terminated at weaning. At this point, immunopathology of immune organs could reasonably be performed, and functional assessment via the TDAR can be made (though response levels are not as robust as in the adult) (Ladics et al. 2000).

The approaches put forth in Figure 12A and B open the possibility that DIT endpoints might be designed into existing DART protocols. There is no question that the scientific rigor of the stand-alone studies could be significantly improved if scientists could integrate all potential drug-/chemical-mediated effects into a single study, including such endpoints as onset of sexual maturation, neurofunctional assessment, and maternal toxicity, rather than having to interpret data from different studies in which exposures may have been different, where effects may not have been observed, and/or where some endpoints may not have been assessed. The 2002 FDA guidance on immunotoxicity evaluation suggests that DIT can be performed using rats in standard reproductive toxicity studies, rather than needing a stand-alone study. Moreover, although the EPA currently has not published an official guideline/protocol for DIT testing, the Agency has consistently encouraged combination endpoint studies where it makes sense to do so. This approach is also consistent with the recommendations from van Loveren and colleagues (van Loveren et al. 2003; van Loveren and Piersma 2004) and from a number of recent workshops addressing DIT (Holsapple et al. 2005; Ladics et al. 2005; Luster et al. 2003), where scientists emphasized that it would be best to add DIT endpoints to reproductive toxicity studies whenever possible.

A combination DART-DIT-DNT study proposal has also been put forth by the ILSI HESI ACSA Technical Committee (Cooper et al. 2006), which may be broadly applicable to environmental chemicals. These protocols are already required for chemical registration, and by design, address most of the critical windows of development. At the core of the HESI ACSA Life Stages paradigm is an extended one-generation study (Cooper et al. 2006) (Figure 12C ). Selected F₁ generation pups in three sets are treated continuously until PND 70. The Task Force recommended the following: Set 1a for clinical pathology including clinical chemistry (with thyroid hormones), hematology and urinalysis followed by gross necropsy with a determination of a number of relevant organ weights; set 1b for limited histopathology of relevant organs and DNT assessment; set 2 for DIT including an assessment of the T cell–dependent antibody response (TDAR) between PNDs 63 and 70; and set 3 for estrous cycle monitoring and any other triggered endpoints (e.g., on a case-by-case basis, a cellular assay may be considered if the TDAR were negative, and phenotypic evaluation may be considered if the TDAR were positive). Those interested in these flexible modifications are encouraged to consult their respective regulatory authorities. It is important to emphasize here that while there appears to be agreement that DIT (and DNT) endpoints can be integrated into DART protocols for hazard identification purposes, this should not be viewed as an endorsement of these super-studies as the study design of the future. Most importantly there is general agreement by industry and regulatory scientists that flexibility to alter existing testing paradigms to meet the needs to the project (e.g., modify dosing regimens, adding endpoints and dosing arms) is the key message.

DIT Testing in Nonrodent Models

The preceding paragraphs have defined an approach that considers testing in rodents. However, rodents are not always the most appropriate test species for pharmaceuticals. Occasionally primates, and rarely (to date) dogs, have been used to assess DIT. The issues facing DIT in the rodent (e.g., prior published study design issues, endpoints, exposure) can also be applied to DIT testing in dogs or nonhuman primates. Buse et al. (2003) have published a proposed combined embryofetal/peri- and postnatal development study in the nonhuman primate that includes an evaluation of DIT that is depicted in Figure 12D . Like the standard peri- and postnatal design for rodents (Figure 12A), direct dosing of offspring is not performed and thus “recovery” of potential DIT effects can occur postnatally. A formal DIT study design in monkeys has not been vetted throughout the immunotoxicology scientific community, though modifications of this proposed design are being utilized for regulatory purposes with human investigational therapeutics. Buse (2003) notes that there are several “immune function” assays that can be used in monkeys (TDAR, DTH, cytokine analysis, ex vivo blastogenesis, phenotyping, natural killer [NK], macrophage function, etc.). Importantly, the same cautions/considerations that apply to rodents regarding these assays would also apply to the primate. Other groups have also assessed DIT in primates (Neubert et al. 2002; Hendrickx et al. 2005) and beagle dogs (Miller and Benjamin 1985); however, like many earlier studies in rodents, exposure paradigms vary and are often associated with in utero exposure only.

Key Considerations in DIT Study Design and Interpretation

As described earlier in this review, immune system development of mammals varies (e.g., the rodent is delayed compared with humans). Therefore, this should be considered when designing studies (e.g., what is the question) and interpreting the information and its potential relevance to humans. Because it has been demonstrated that sex-specific outcomes can occur for several developmental immunotoxicants, the more robust assessment would include an evaluation of both genders. It has also been recognized that an understanding of offspring exposure during gestation, lactation, and juvenile life stages is important. This includes an understanding of placental transfer and secretion into breast milk. A negative DIT study (no adverse effects) in the absence of confirming exposure to the offspring is not sufficient, and overexposure resulting from poor high-dose selection confounds the ability to determine whether effects are direct or secondary to overt toxicity. To date, exposure has not been effectively characterized in DIT studies. As an initial screen, the best approach may be to assess all of the critical windows at once by assuring exposure to the developing organism occurs in utero, during lactation, and during juvenile/adolescent development, which may require direct dosing to begin soon after birth. Toxicokinetic data can and should be used wherever possible for designing protocols and in determining the need for direct dosing to preweaning offspring (Holsapple et al. 2005). Considerations for direct dosing have been deliberated elsewhere (Moser et al. 2005). Similarly, doses resulting in maternal toxicity can impact immunotoxicology endpoints (Carney et al. 2004) and this should be considered during study design. Finally, consideration should be made for including an examination of reversibility into any DIT assessment. There is no consensus on an appropriate duration, though 2 to 4 weeks has been suggested (Holsapple et al. 2003).

RISK ASSESSMENT

To understand where DIT fits in the risk assessment process, the process itself needs to be defined. Characterization of the risk of a specific adverse event occurring involves hazard identification/characterization, dose-response evaluation, and exposure assessment in the population of interest. Following that, a most sensitive endpoint in the most sensitive species is identified and up to five different uncertainty factors (UFs) are applied (interspecies, intraspecies, subchronic to chronic, LOAEL to NOAEL, and completeness of the database) to account for up to 10-fold differences in order to derive an exposure that is considered safe. As a result of the passage of the Food Quality Protection Act in 1996, an additional 10-fold safety factor (“the FQPA factor”) was mandated for food-use pesticides, when deemed necessary, based on the available toxicity and exposure information because of specific concerns regarding protection of children’s health. This factor is used to address residual uncertainties not accounted for through the use of the five other uncertainty factors (EPA 2002b). Though risk assessment guidelines have been developed by EPA for developmental and reproductive toxicity, neurotoxicity, and carcinogenicity (EPA 1991, 1996, 1998, 2005), there are as yet no similar guidelines for immunotoxicity to account for effects on either the developing or the adult immune systems. Of note, however, the derivation of the reference dose (RfD) for tributyltin relied on the effects of the chemical on the adult immune system as the most sensitive indicator of toxicity (EPA 1997).

Kimmel et al. (2005) describe how developmental immunotoxicity data may be used in risk assessment for environmental chemicals. EPA considers the definition of developmental toxicity to be “… the study of adverse effects on the developing organism that may result from exposure prior to conception (either parent), during prenatal development, or postnatally to the time of sexual maturation. Adverse developmental effects may be detected at any point in the lifespan of the organism. The major manifestations of developmental toxicity include (1) death of the developing organism, (2) structural abnormality, (3) altered growth, and (4) functional deficiency” (EPA 1991). Considering the broad nature, DIT clearly fits into this definition as the DIT endpoints evaluated may demonstrate a functional deficiency detected at some point in the life of the animal. DIT regulatory testing of environmental chemicals is expected to provide both hazard identification and dose-response information that can be used in formulating a risk assessment for the chemical. Additionally, Kimmel et al. propose that when there is evidence in the data being reviewed for any given chemical that the immune system of the adult organism may be a target tissue, and if there are no data available on DIT, the risk assessor could consider the application of an uncertainty factor to address the perceived deficiency in the database. Although this may seem even more conservative, in truth this is simply the normal risk assessment process for any adverse effect with a consideration of whether the database warrants the application of a data base uncertainty factor. Formal risk assessment calculations (reference values, benchmark dose levels, etc.) are not performed to the same didactic stringency for pharmaceuticals primarily because, unlike exposure to environmental chemicals, there is the expectation of “benefit” of the therapeutic to the patient population and generally, the database is relatively well defined. Still, risk characterization is required to determine the risk-benefit ratio and evaluate the acceptability of the identified risk(s).

DIT is still evolving as a science. There is no question that we have learned a lot in recent years about the developing immune system of individual species, but we currently do not really know how to translate the significance of drug or chemical effects on these systems to potential human hazards (e.g., exposure assessment and epidemiology). At just what point—for any endpoint selected—does an alteration in the developing (or adult) immune system become “adverse”? The question is particularly relevant when the immunological effects, as might be expected to occur from inadvertent exposures, are minimal-to-moderate in nature. Epidemiological evidence exists that indicates that the normal functioning of the immune system of children can be disrupted by alterations that occur to the developing organism (e.g., heritable traits resulting in moderate to severe immunodeficiency). It has also been suggested that it may be feasible to incorporate an assessment of DIT in human clinical trials (Neubert et al. 2002) in order to begin gathering these critical human data, though to date there is no consensus on translatable biomarkers. Similarly, it is evident that exposure to both drug and nondrug chemicals can have adverse consequences on the developing immune system. However, as emphasized earlier, the study designs that resulted in these nonclinical data were highly variable and often only evaluated certain critical developmental windows. Studies to examine the true persistence of DIT effects throughout the life of an animal have generally not been performed, or made publicly available. Indeed, the likelihood that the majority of developmental immunotoxicants would produce profoundly persistent or permanent effects is probably not high. Nevertheless, it is important to consider that discrete events resulting in only temporary changes can still have significant health effects in an age-specific population (e.g., reduced response to childhood vaccination). The state of the science of DIT is on the right track with respect to the future identification of developmental immunotoxicants for use in regulatory risk assessment, and product development decision making by both industry and regulators.

CONCLUSIONS AND OPPORTUNITIES FOR THE FUTURE

With regards to the development of testing strategies for DIT, although no one can deny or challenge the importance of maintaining children’s health, the implementation of guideline tests or guidance documents to address the potential for DIT can only be based on a solid scientific foundation. A number of regulatory bodies have considered recommendations for evaluating the developing immune system. However, despite this attention, research regarding DIT and the appropriate methods to assess it has been limited. Extensive evaluation and validation of several assays/endpoints has been performed in adult rodents; however, there are few data available regarding whether these assays/endpoints that have been optimized in adult animals could be utilized or would be sufficiently sensitive in assessing the functionally immature immune system of neonatal or juvenile animals. In one example, an antibody response to sRBCs of sufficient magnitude was observed with the antibody forming cells (AFC) assay in rat weanlings, but not rat pups (Ladics et al. 2000). Some endpoints (e.g., analysis of immune cell subtypes by flow cytometry or total serum immunoglobulin levels) have been shown to lack sensitivity in adults and would therefore likely lack sensitivity in assessing DIT. In addition, there is a need for a validated assay that truly measures cell-mediated immunity for assessing the potential DIT of a xenobiotic (Holsapple 2002; Luster et al. 2003; Holsapple et al. 2005). The utility, sensitivity, and specificity of other endpoints, such as the measurement of various cytokines or the integration of ‘-omic’ techniques, for assessing DIT or immunotoxicology in general are in the early stages of evaluation. Before such endpoints can be implemented, it is important to establish a database(s) of baseline values for normal levels of cytokines or gene expression that occur in immune tissues and organs during the various windows of immune system development. In addition, it is important to understand how the cytokine or gene expression profile observed relates to the administered dose and the kinetics of the xenobiotic-induced response (Burns-Naas et al. 2006; Luebke et al. 2006). Regardless of the endpoints that are established to evaluate DIT, it has been recommended that when possible, DIT assays/endpoints be added onto existing developmental and reproductive toxicology protocols in order for all xenobiotic-induced effects to be integrated into a single study design to facilitate data interpretation (Ladics et al. 2005). Although various forums have indicated the need for validated methods to evaluate DIT, to date research efforts regarding DIT and the appropriate methods to assess it have been limited. Additional resources are needed to establish DIT screening endpoints to further investigate whether the developing immune system may be at greater risk to certain xenobiotics compared to adults.

One point that is interesting to consider is whether there are developmental immunotoxicants that do not exert immunotoxicity in adult animals. As would be expected at this stage in its evolution, the chemicals used to date to investigate developmental immunotoxicity are known immunotoxicants in adult animals—TCDD, lead, immunomodulators like cyclosporine, or cytotoxic agents such as cyclophosphamide. For human therapeutics, if the present testing paradigms remain in place (e.g., trigger-based assessment), it is unlikely that we will ever obtain an answer from this sector. This information is more likely to be acquired through regulatory-based testing of environmental chemicals (e.g., pesticides).

Ultimately, the most appropriate model for evaluating DIT in humans is, of course, humans. Our ability to do this presently is an area that needs further research (Luster et al. 2005). The use of T-cell receptor rearrangement excision circles (TRECs), immunophenotyping, the evaluation of cytokines and serum immunoglobulin (Ig), and the quantification of the immune response to childhood vaccines have been suggested as possible endpoints or biomarkers. None of these has been evaluated to an extent that reasonable recommendations can be made, and some (serum Ig and phenotyping) have been considered relatively insensitive markers in the adult and thus may also have questionable value for assessing the developing immune system. More information is needed to determine biomarkers of adverse DIT events and whether an effect on a biomarker from an adult study can be applied to set safe levels in juveniles. And despite the available animal data to suggest differential sensitivities between adult and juveniles, it has not been unequivocally demonstrated that juvenile humans are truly more sensitive to immune system perturbations than adults. Additionally, as mentioned previously, the use of immunotoxicology data in risk assessment is made more difficult because we do not have a full appreciation of how much of a change in a single (or multiple) endpoint is actually “adverse” to the organism, and we must still apply suppositions to account for differences in the overall immunocompetence of an individual, which is affected by genetics, age of the child (or adult), gender, use of certain medications, nutritional status, and environment. Along these lines, Holsapple et al. (2007) have suggested that a scientifically sound framework be established that allows for more accurate and quantitative interpretation of such data in the risk assessment process is needed, and may require, for example, development of a model to equate moderate changes in leukocyte counts, CD4 cell numbers, and/or immunoglobulin levels, tests that can be readily performed in children, to potential changes in the incidence or severity of infectious diseases.

Finally, the obvious issue yet to be addressed remains the impact of DIT on the potential acquisition of immune-mediated disease (e.g., hypersensitivity and autoimmunity) in early life stages. Even in adults there are no models characterized that adequately address systemic hypersensitivity or the induction of autoimmunity and these have certainly not been addressed in juvenile animals. However, Holsapple et al. (2007) did note that it was important to consider that the methods used to assess contact sensitivity in adult animals do, in fact, measure a DTH response, a cellular response that has been recognized as an integral part of the robust assessment of DIT. Might this be a point at which we can begin to assess the other side of the immune function coin in developing animals?